An Analysis of the Construct-Related and Ecological Validity of the BNI Screen for Higher Cerebral Functions by Peter J. Wass: A Synopsis
George P. Prigatano, PhD
Division of Neurology, Barrow Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona
A synopsis of Peter J. Wass’ dissertation findings on the BNI Screen (BNIS) for Higher Cerebral Functions is presented. The BNIS Total score correlates with numerous neuropsychological measures. Moreover, the various subtest scores correlate with other tests purporting to measures similar abilities. Finally, the BNIS correlates with the commonly used measure of functional outcome at time of rehabilitation discharge after brain injury. Collectively, the findings support the ecological and construct validity of the BNIS.
Key Words: BNI Screen, cross-validation, Canadian population, construct validity, ecological validity
In 1997 Peter J. Wass conducted his doctoral dissertation at the University of Windsor, Windsor, Ontario, Canada on the construct and ecological validity of the BNI Screen (BNIS) for Higher Cerebral Functions. This article summarizes key findings from Dr. Wass’ dissertation and was prepared with his permission and approval.
Materials and Methods
Eighty-five patients (49 males, 29 females) who were discharged from the Alberta Hospital Ponoka Brain Injury Rehabilitation Program in central Alberta served as subjects. All patients had a traumatic brain injury. Data were accumulated from 1994 to 1996.
The mean age of patients at the time of examination was 36.5 years (standard deviation, (SD)=13.0; range, 17 to 66). Their mean education level was 11.7 years (SD=2.1;range,6 to 18 years).
Sixteen percent were tested within 3 months of their brain injury. Forty-eight percent were tested 3 to 12 months after their injury. Thirty-six percent were tested after 12 months after their brain injury.
Each subject was administered the BNIS and a battery of neuropsychological tests that included the Wechsler Adult Intelligence Scale-Revised Form, selected subtests from the Wechsler Memory Scale-Revised Form, the Controlled Oral Word Association subtest of the Multilingual Aphasia Examination, the Wisconsin Card SortingTest, and selected measures of the Halstead-Reitan Neuropsychological Test battery. Those tests included the Halstead Category Test, the Trail-Making Test (Parts A and B), and the Groove Pegboard Test.
Each patient also was administered the Functional Independence Measure (FIM)1 and the Functional Assessment Measure (FAM). The latter test was used to expand the FIM and to obtain information about cognitive functioning as part of the Uniform Data System for medical rehabilitation.
Various correlational analyses were performed. The correlations assessed the predictive relationships between various subtests from the BNIS and existing neuropsychological tests and measures of functional performance at discharge. The goals of the data analysis were to assess the relationship of performance on the BNIS to various neuropsychological measures and to determine its construct and ecological validity.
Most of the correlations between the BNIS Total score and subtest scores with commonly used neuropsychological measures were statistically significant, although magnitudes ranged from mild to moderate (Table 1). The item dealing with awareness did not correlate with any of the neuropsychological measures. The important theoretical implications of this finding are discussed below.
The BNIS Total score correlated the strongest with the Digit Symbol subtest of the Wechsler Adult Intelligence Scale-Revised (WAIS-R) Form. This particular subtest of the WAIS-R has been reported to be sensitive to brain dysfunction. The Speech and Language subtests of the BNIS consistently correlated with Vocabulary and knowledge of verbal information regarding the world, adding to its construct validation. The Memory subtest of the BNIS correlated significantly with Logical Memory and Visual Reproduction scores obtained on the Wechsler Memory Scale-Revised Form. It also correlated strongly with the Digit Symbol subtest, which has a memory component. The Visuospatial Problem-Solving subtest of the BNIS correlated strongly with a number of neuropsychological measures. The strongest correlations were observed with the Digit Symbol and Block Design subtests. These findings further support the construct validity of the BNIS and collectively suggest that the BNIS is measuring what it attempts to measure (i.e., has good construct validity).
An important question is how neuropsychological test findings correlate with basic functioning in the “real world.” There was a strong, positive correlation between the total FIM + FAM score and the BNIS Total score (r =+.65; p<.001, Table 2). These data support the ecological validity of the BNIS. At time of discharge, the BNIS Total score has predicted functional independence from a postacute brain injury rehabilitation program.
Dr. Wass’ dissertation analyzed the construct and ecological validity of the BNIS. This synopsis of his research highlights two points. First, the BNIS Total score and its various subtests correlate with known neuropsychological measures. The only exception was the awareness measure, which consisted of one item that requires subjects to predict the number of words that they will recall after a brief distraction.
The failure of this measure to correlate with neuropsychological measures can be explained on the basis of at least two factors. First, the reliability of the awareness measure may be weak because it consists of a single item. Second and more probable, the ability to accurately perceive one’s abilities may be independent of traditional measures of memory and problem solving. Prigatano and Wong suggested that human self-awareness is an emergent brain function mediated by many higher brain functions. It represents an integration of thinking and feeling. Consequently, it is not assessed by standard IQ or memory measures. The lack of correlation between this measure and traditional neuropsychological measures adds some support to this theoretical notion.
That the BNIS Total score correlated with almost every neuropsychological measures is impressive. This finding provides strong support for the construct validity of the screen. Equally important to rehabilitationists is the observation that the BNI Total score is strongly correlated with the FIM+FAM total score. Clinicians frequently need a brief assessment that allows them to make statements about a patient’s ability to function independently. The Wass study indicates that the BNIS can be used for this purpose, at least in postacute brain dysfunctional patients. Prigatano and Wong have shown that the BNIS Total score correlates equally strongly with the achievement of rehabilitation goals after acute inpatient neurorehabilitation.
- Crewe NM, Dijkers M: Functional assessment, in Cushman LA, Scherer MJ (eds): Psychological Assessment in Medical Rehabilitation. Washington, DC: American Psychological Association, 1995, pp 101-144
- Heaton RK: A Manual for the Wisconsin Card Sorting Test. Odessa, FL: Psychological Assessment Resources, Inc., 1981
- Prigatano GP, Wong JL: Cognitive and affective improvement in brain dysfunctional patients who achieve inpatient rehabilitation goals. Arch Phys Med Rehabil 80:77-84, 1999
- Reitan RM: Theoretical and methodological bases of the Halstead-Reitan Neuropsychological Test Battery, in Grant I, Adams KM (eds): Neuropsychological Assessment of Neuropsychiatric Disorders. New York: Oxford University, 1986, pp 3-30
- Reitan RM, Wolfson D: The Halstead-Reitan Neuropsychological Text Battery: Theory and Clinical Interpretation. Tucson, AZ: Neuropsychology Press, 1985
- Wass PJ: An Analysis of the Construct-Related and Ecological Validity of the Barrow Neurological Institute Screen for Higher Cerebral Functions (PhD dissertation). Windsor, Ontario: University of Windsor, 1997
- Wechsler D: Wechsler Adult Intelligence Scale- Revised. San Antonio, TX: The Psychological Corporation, 1981
- Wechsler D: The Wechsler Memory Scale-Revised. San Antonio, TX: The Psychological Corporation, 1987